What’s the chance of...the away team winning?

Coherent probability surfaces for agents that make trade-off decisions.

Lots of places enable guesses about the same event. They don’t always agree.

ChanceDB combines those predictions into one clean, standard format that machines can read, trust, and act on.

Rules check that surface. Agents act from it.

Alpha previewLive and evolving. APIs may change.

Live surfaces

Discrete events. Continuous ranges. Same canonical surface.

Arsenal vs Tottenham
Terminal Winner · Cross-Venue Surface
View Snapshot
Venue inputs
Polymarket
Home 0.50
Draw 0.22
Away 0.28
Kalshi
Home 0.56
Draw 0.24
Away 0.20
Canonical surface
Home
0.53
Draw
0.23
Away
0.24
Aggregate overround: 1.05
Arbitrage detected: +500 bps
BTC Hits Price · 2026-12-31
Canonical Probability Surface
Polymarket + Kalshi reconciled
View Surface
Venue inputs
Polymarket
Median 92k
Implied skew: positive
Kalshi
Median 98k
Wider upper tail
Continuous probability distribution
p = 0.10Q(p) = $51,200p = 0.50Q(p) = $94,800p = 0.85Q(p) = $142,600Full inverse CDF available via API
Two surface types. One surface architecture.

Discover → Compose → Route

Three steps. One surface.

Discover

Browse surfaces

Find existing surfaces across domains.

Explore Surfaces
Compose

Integrate sources

Combine markets, models, and experts.

Open Studio
Route

Define policy logic

Route surface changes to agents or webhooks.

Open Routing

How it works

Forecast data

Markets · Models · Experts

Surface

Normalized probability · Regime metrics

Policy

Thresholds · Routing logic

Action

Agents · Webhooks · Publications

Relationships

See how outcomes move together
When one thing shifts, what else moves? Model connected scenarios across markets.
Coming soon